Green tea and brain health: what the evidence actually supports

Short answer: There is a real, repeatedly observed association between drinking green tea and better brain ageing — and it is smaller than the headlines suggest. Around three cups a day is where the studies cluster, and the best-measured benefit so far is roughly 3% less white-matter damage on brain scans. Every study behind that is observational, which means it can show a pattern but cannot prove green tea caused it. Almost all of it comes from Japan and China, and when researchers looked at European populations separately, the association wasn't statistically significant. The most useful thing to know isn't the size of the benefit at all — it's the difference between drinking the tea (safe, mildly promising) and taking green tea extract supplements (a well-established cause of liver injury). If you take one thing from this page, take that one.

This is educational, not medical advice. Green tea interacts with some medications — including blood thinners — and concentrated extracts carry a documented risk of liver injury. Talk to your doctor or pharmacist before adding a supplement, and if you're worried about memory changes that are new, worsening, and affecting daily life, that warrants a clinical appointment.

What the newest research actually found

Two Japanese studies published in 2025 are behind most of the coverage you'll have seen.

The brain-scan study. Researchers at Kanazawa University looked at 8,766 older adults without dementia who had MRI scans as part of the Japan Prospective Studies Collaboration for Aging and Dementia, and compared their brain images against how much green tea they reported drinking.

The finding: people drinking about 600 mL a day (roughly three cups) had around 3% less white-matter lesion volume than people drinking 200 mL or less. At 1,500 mL a day — seven or eight cups — the difference was about 6%.

White-matter lesions are small areas of damage to the brain's wiring, mostly caused by the health of tiny blood vessels. They accumulate with age, and more of them is associated with faster cognitive decline and higher dementia risk. Fewer is better.

Three things in the same study matter just as much and get quoted far less:

• Green tea showed no association with hippocampal volume or total brain volume — the benefit, if it is one, looks specific to blood-vessel-related damage rather than general brain shrinkage.
• Coffee showed no significant association with any of the three measures in this study.
• The association was not seen in participants who had depression or carried an APOE ε4 gene variant.

That last point deserves a moment. Subgroup findings are statistically fragile and shouldn't be treated as a rule. But it's an honest signal that whatever is going on here isn't universal — and it is missing from essentially every news write-up of this study.

The authors were careful about what they had: this was a cross-sectional study — a single snapshot, not a follow-up over time — and they state directly that the tea components involved weren't experimentally validated as causal.

(Shibata S, Noguchi-Shinohara M, Shima A, et al., npj Science of Food 9, 2, 7 January 2025; summarised by Kanazawa University)

The follow-up study. A separate 2025 paper followed 13,660 Japanese adults aged 40–74 for a mean of 11.5 years, during which 682 developed dementia. People in the highest green-tea group (600 mL a day or more) had a 25% lower rate of dementia than those in the lowest group (under 94 mL a day) — a hazard ratio of 0.75, with a significant trend across the groups. Each extra 150 mL cup per day was associated with roughly a 5% lower rate.

This design is stronger than a snapshot, because tea drinking was recorded before anyone developed dementia. But the authors flag a real problem: they had no baseline cognitive testing. If some participants already had very early, undetected cognitive changes that made them drink less tea, the association would look larger than it is. That's reverse causation, and it haunts almost every diet-and-dementia study.

The sex results were also unstable — the trend was clearly significant in neither group (men p = 0.18; women p = 0.05), and an earlier Japanese study found the clearest effect in men, the opposite way round. When a signal flips direction between cohorts, the honest read is that it isn't reliable yet.

(Kaise R, et al., The Journal of Nutrition, Health & Aging, June 2025)

How many cups? The number the studies converge on

Across the research, the answer keeps landing in the same place: about three cups a day, and more is not better.

A 2025 meta-analysis pooled 18 studies covering 58,929 people and found green tea consumption associated with lower odds of cognitive impairment overall (OR 0.63), of dementia (OR 0.74) and of mild cognitive impairment (OR 0.64). For Alzheimer's disease specifically, the estimate pointed the same way but was not statistically significant (OR 0.57, CI 0.21–1.59) — the confidence interval is wide enough to include no effect at all.

The dose pattern was informative: low intake showed no significant association, moderate intake did (OR 0.73), and high intake was similar (OR 0.64). There's a floor, and then it flattens.

(Zhou S, Zhu Y, Ren N, Wu M, Liu Y, Neuroepidemiology 2026;60(1):178–199, published online 13 February 2025)

The largest study outside East Asia points the same way about the ceiling. In 377,592 UK Biobank participants followed a median of nine years, tea intake had a U-shape: moderate drinkers (roughly one to six cups a day) had about a 14–20% lower rate of dementia than non-drinkers, with the low point around three cups — and above six cups a day the benefit disappeared.

One important caveat, because it's the sort of thing that gets lost: that study didn't record what kind of tea people drank, and UK tea drinking is overwhelmingly black tea. It tells you something about the shape of the dose curve. It is not green-tea evidence.

(Hu HY, Wu BS, Ou YN, et al., Translational Psychiatry, 26 April 2022)

The problem nobody puts in the headline

Here is the finding that should reset how you read all of the above.

When the 2025 meta-analysis broke its results down by population, the association was strong in Chinese cohorts (OR 0.55) and present in Japanese cohorts (OR 0.74) — and in European populations it was not statistically significant.

That is not a small footnote. It means the evidence base for green tea and brain health is almost entirely East Asian, drawn from places where green tea is a daily cultural staple. And in those settings, the person who drinks six cups of green tea a day is systematically different from the person who drinks none: different diet, different activity, often different income and education. Researchers adjust for what they can measure, but they cannot adjust for everything, and this is exactly the situation where residual confounding is hardest to rule out.

The authors also reported high heterogeneity between studies (I² = 77.6%) and a statistical test suggesting possible publication bias — studies finding nothing are less likely to be written up.

There's a simpler sanity check too, and the Fisher Center makes it well: dementia rates are not lower in Japan or China, the countries where green tea consumption is highest in the world. If green tea were a powerful lever, you'd expect to see it at the population level. We don't.

None of this means green tea does nothing. Green tea contains catechins (particularly EGCG), L-theanine and caffeine, and there are plausible mechanisms — antioxidant and anti-inflammatory effects on small blood vessels — that fit the white-matter finding neatly. It means the honest summary is: a small, plausible, consistently-observed association that has never been tested in a trial designed to answer the question.

Drink the tea, skip the pills

This is the part with real-world stakes, and it's absent from every page currently ranking for this topic.

Brewed green tea has not been associated with liver injury. The US National Institutes of Health's LiverTox database notes the opposite — regular green tea drinking is associated with lower liver enzyme levels.

Concentrated green tea extract is a different substance entirely. LiverTox assigns it a likelihood score of A — "well established cause of clinically apparent liver injury" — with more than 100 published cases of liver damage attributed to it. European food-safety regulators concluded that extract doses equivalent to 800 mg of EGCG or more per day can be associated with significant liver-enzyme elevations. For scale, a brewed cup contains somewhere around 25–200 mg of catechins in total.

And the mechanism is not the reassuring kind. In people who developed liver injury from green tea extract, 72% carried the HLA-B*35:01 genetic variant, versus 11–15% of controls — meaning the reaction is largely immune-mediated and idiosyncratic. It isn't a straightforward "too much is bad" dose effect you can sense building. Some people react and most don't, and there's no practical way to know in advance.

(LiverTox: Green Tea, NCBI Bookshelf, National Institute of Diabetes and Digestive and Kidney Diseases)

This maps onto a broader pattern. In its 2026 risk-reduction guidelines, the World Health Organization recommends against supplements — including vitamin B, vitamin E and multivitamins — taken solely to reduce dementia risk. We covered that in what the WHO now says about lowering dementia risk, and it echoes what we found looking at brain supplements and turmeric: the food keeps looking better than the pill, and the pill occasionally carries risks the food does not.

Three practical points worth raising with a pharmacist: green tea can interact with anticoagulant medications including warfarin; it can interfere with iron and folate absorption (drinking it between meals rather than with them reduces this); and at roughly 30–60 mg of caffeine a cup, seven or eight cups is a real caffeine load — given how strongly sleep is tied to dementia risk, trading sleep for tea would be a bad exchange.

Reviewing the same literature, the Alzheimer's Drug Discovery Foundation puts the studied safe range at three to five cups a day, up to about 1,200 mL — comfortably where the benefit signal sits anyway.

Where green tea sits among things that actually move the needle

For a sense of proportion: the 2024 Lancet Commission identified 14 modifiable risk factors together associated with around 45% of dementia cases worldwide. Green tea is not one of them. Hearing loss, high blood pressure, smoking, physical inactivity, diabetes, depression, social isolation, air pollution, high LDL cholesterol and untreated vision loss are.

So green tea belongs in the same category as the rest of the MIND diet — a dietary pattern with modest but real support — and nowhere near the same league as treating blood pressure, getting your hearing checked or moving regularly. If you're swapping green tea in for sugary drinks, the swap itself is probably the bigger win.

So what should you do?

1. If you like green tea, drink it — around three cups a day, where the evidence clusters. It's pleasant, cheap and safe, and the downside is essentially nil.
2. Don't force it. The effect sizes are small and unproven; nothing here justifies drinking something you dislike.
3. Don't take the extract. It's a documented cause of liver injury with no evidence of extra brain benefit. If you're already taking one, mention it to your doctor.
4. Watch the caffeine and the timing — not with iron-rich meals if you're low on iron, not late enough to cost you sleep.
5. Spend most of your effort elsewhere. Blood pressure, hearing, activity, blood sugar, smoking and sleep are the factors worth building a routine around.

Find out which factors matter most for you

Green tea is a small lever. Knowing which of the big ones apply to you is a better use of ten minutes.

For the wider picture, start with can you reduce your dementia risk? or check your dementia risk.

How we research and review our content →

This article is educational and is not medical advice, diagnosis, or treatment. The studies cited describe findings in general populations, not any individual. Solenna is not a medical device and does not diagnose, prevent, treat, or cure Alzheimer's disease or any other condition. Risk reduction means lowering probability, not eliminating it; individual results vary and no outcome is guaranteed. Green tea extract supplements are a different product from brewed tea and carry documented liver-injury risk; discuss any supplement with your clinician or pharmacist, particularly alongside prescription medication.